PET Spatial Resolution and Positron Range

Q: What limits spatial resolution in PET?

Four physical factors dominate: the finite size of the detector element (about half the crystal width at the center of the field of view), annihilation photon non-collinearity (the two 511 keV photons are not exactly 180° apart), positron range (the distance the positron travels before annihilating), and detector penetration/decoding errors. Reconstruction and sampling add further blurring. These combine in quadrature and set a floor below which finer detectors cannot push resolution.

Q: What is positron range and why does it blur PET images?

A positron does not annihilate at the point of decay. It travels a short, energy-dependent distance through tissue, scattering and losing energy, before it meets an electron and annihilates. PET reconstructs the annihilation location, not the decay location, so the separation between them blurs the image. The higher the positron's energy, the farther it travels and the greater the blur.

Q: Why is Ga-68 or Rb-82 imaging blurrier than F-18?

Positron energy differs sharply between radionuclides. F-18 emits low-energy positrons (0.63 MeV maximum) that travel well under a millimeter on average, while Ga-68 (1.90 MeV) and Rb-82 (3.38 MeV) emit high-energy positrons that travel several millimeters. On the same scanner, the higher positron range of Ga-68 and Rb-82 measurably degrades spatial resolution and small-lesion recovery.

Q: What spatial resolution do modern PET/CT scanners achieve?

Modern digital PET/CT systems measured under NEMA NU 2-2018 with an F-18 point source typically report transverse resolution near 3.8–3.9 mm FWHM at a 1 cm radial offset, and total-body long-axial-FOV systems reach about 3.0 mm near the center. Because the standard uses F-18, these numbers understate the blur seen clinically with higher-energy emitters.

Q: Does non-collinearity depend on scanner size?

Yes. The annihilation photons deviate from a perfect 180° by about 0.5° FWHM because the positron-electron pair retains a small residual momentum. This angular uncertainty translates into a positional blur proportional to the detector-ring diameter, so larger-bore scanners suffer more non-collinearity blur than small-bore or preclinical systems.

Q: Can positron range blurring be corrected?

Partly. Resolution-recovery and positron-range corrections during reconstruction can model the blurring kernel, but many use a generic F-18 model that undercorrects high-energy emitters such as Ga-68. Accurate correction requires a radionuclide-specific range kernel, and residual blur remains because the range distribution is broad and tissue-dependent.

Q: How does spatial resolution affect SUV and small lesions?

Limited resolution causes the partial volume effect: activity from a small lesion spills into surrounding voxels, so a lesion smaller than about twice the FWHM measures a falsely low SUV and can be missed entirely. Better resolution improves both detection of small lesions and the quantitative accuracy of SUV measurements.

PET spatial resolution is not primarily a software problem — it is bounded by physics. Three effects set the floor: the finite size of the detector element, the roughly 0.5° non-collinearity of the two annihilation photons, and the distance a positron travels before it annihilates. Of these, positron range is the one that changes most from one radionuclide to another, which is why an F-18 FDG image is visibly sharper than a Ga-68 PSMA or Rb-82 cardiac image acquired on the very same scanner.¹²

This article works through the physics of PET spatial resolution, quantifies each contribution, tabulates positron range across the common radionuclides, and connects the numbers to what a nuclear medicine physicist and clinician actually see. DRPS supports this work through its PET/CT and nuclear medicine physics and accreditation support services across Florida, Maryland, Virginia, Washington DC, California, and beyond.

Introduction

Ask what determines how small a lesion PET can resolve and the intuitive answer — "the detector" — is only one-quarter of the story. Even a hypothetical PET scanner built from infinitely small detectors could not resolve a point source perfectly, because the physics of positron decay and annihilation blur the signal before a single photon reaches a crystal. The measured image is a convolution of several blurring processes, and understanding them explains why resolution has improved only modestly over decades of detector engineering, why the choice of radionuclide matters, and why quantitative accuracy (SUV) depends on physics the operator cannot change.¹

The central insight is that PET localizes an annihilation, not a decay. Between those two events, the positron travels. That single fact — positron range — is the difference between a crisp F-18 study and a softer Ga-68 one, and it is the thread that runs through this article.²³

Topic Explanation

What PET actually measures

A positron-emitting nucleus decays and ejects a positron. The positron scatters through surrounding tissue, shedding kinetic energy over a tortuous path, until it slows enough to combine with an electron. The pair annihilates, converting their mass into two 511 keV photons emitted almost exactly back-to-back. PET detects that photon pair in coincidence and assigns the event to the line connecting the two detectors — the line of response.

Two physical realities blur this scheme immediately:

The annihilation happens where the positron stopped, not where the nucleus decayed. The offset between them is the positron range.
The two photons are not emitted at exactly 180°. Because the annihilating pair retains a small residual momentum, the photons deviate from collinearity by about 0.5° FWHM — the non-collinearity effect.

Add the finite detector element size and imperfect event decoding, and you have the four fundamental contributors to PET blur.¹

The resolution budget

The blurring processes are approximately independent, so their FWHM contributions add in quadrature (as for convolved Gaussians). A widely used expression for the reconstructed system resolution of a PET scanner combines them as:¹

Γ = 1.25 (\frac{d}{2})^{2} + b^{2} + (0.0022 D)^{2} + r^{2}

where:

$\frac{d}{2}$ is the detector-element contribution ( $d$ = crystal width); a point source midway between opposed detectors is localized to about half the crystal width.
$b$ is the block decoding error, the uncertainty in identifying which crystal in a block detector fired (near zero for one-to-one silicon-photomultiplier readout).
$0.0022 D$ is the non-collinearity blur, proportional to the detector-ring diameter $D$ (in mm); the constant encodes the 0.5° angular spread.
$r$ is the positron-range contribution, the effective FWHM of the annihilation-point distribution.
$1.25$ is an empirical factor accounting for the additional blurring introduced by filtered-backprojection-style image reconstruction.

The quadrature form has a crucial consequence: the largest term dominates, and shrinking an already-small term barely helps. Once detectors reach ~3–4 mm, further miniaturization yields diminishing returns because non-collinearity and (for high-energy emitters) positron range become comparable in size.¹²

Key Technical Principles

Positron range across radionuclides

Positron range depends on the positron's kinetic energy spectrum, which is set by the decay. F-18 emits soft positrons; Ga-68 and especially Rb-82 emit energetic ones. The table gives representative maximum and mean positron energies and ranges in water, with the qualitative resolution impact.²³⁴

Radionuclide	β⁺ max energy (MeV)	Mean β⁺ energy (MeV)	Mean range in water (mm)	Max range in water (mm)	Relative resolution impact
F-18	0.63	~0.25	~0.6	~2.4	Minimal
C-11	0.96	~0.39	~1.1	~4.1	Small
Ga-68	1.90	~0.84	~2.9	~9	Substantial
Rb-82	3.38	~1.5	~5.9	~15.6	Large

Values are representative literature figures for positron energy and range in water and vary with the tabulation and tissue assumed; they illustrate the ordering F-18 ≪ C-11 < Ga-68 < Rb-82 rather than exact scanner performance.²³

Two nuances matter for interpretation:

The effective resolution blur is smaller than the maximum range. The annihilation-point distribution is sharply peaked (cusp-like) near the origin with long tails, so its FWHM is much less than $R_{ma x}$ . This is why Ga-68 does not blur by 9 mm — the effective FWHM contribution is on the order of a couple of millimeters — but it is still large enough to matter.²⁴
Some radionuclides carry extra baggage. Higher-energy emitters such as Ga-68 and Rb-82 are frequently accompanied by prompt gamma emissions and, for Rb-82, a very high-energy positron branch; these complicate both correction and quantification beyond range alone.³

Non-collinearity scales with bore size

The non-collinearity term, $0.0022 D$ , is fixed by the annihilation physics but scaled by geometry. For a clinical whole-body scanner with a detector-ring diameter of about $D = 800$ mm:

0.0022 \times 800 = 1.76 mm

For a small-bore preclinical scanner ( $D = 200$ mm), the same effect contributes only $0.0022 \times 200 = 0.44$ mm. This is why preclinical PET can chase sub-millimeter resolution while clinical PET cannot — and why positron range, nearly negligible for F-18 in a large clinical bore, becomes the dominant blur in a small preclinical ring imaging Ga-68.¹²

Worked example: F-18 versus Ga-68 on a clinical scanner

Take a modern clinical scanner with $D = 800$ mm, a crystal width $d = 4$ mm, one-to-one SiPM readout ( $b \approx 0$ ), and compare F-18 ( $r \approx 0.5$ mm) with Ga-68 ( $r \approx 2.4$ mm effective).

For F-18:

Γ_{F-18} = 1.25 2^{2} + 0^{2} + 1.7 6^{2} + 0. 5^{2} = 1.25 4 + 3.10 + 0.25 = 1.25 7.35 \approx 3.4 mm

For Ga-68:

Γ_{Ga-68} = 1.25 2^{2} + 0^{2} + 1.7 6^{2} + 2. 4^{2} = 1.25 4 + 3.10 + 5.76 = 1.25 12.86 \approx 4.5 mm

The same hardware delivers ~3.4 mm for F-18 and ~4.5 mm for Ga-68 — a real, physics-driven ~1 mm penalty that no reconstruction fully removes. This ordering is consistent with dedicated positron-range studies: on a benchmark preclinical scanner, measured FWHM from hot capillaries was ~1.4 mm for F-18 but ~3.0 mm for Ga-68, a gap that widens precisely because the small bore removes competing blur and lets positron range dominate.⁴

What NEMA-measured resolution does and does not tell you

Manufacturers report spatial resolution under the NEMA NU 2-2018 standard, which specifies point-source measurements — and, importantly, uses F-18.⁵ Representative modern figures:

A digital 30-cm axial-FOV Discovery MI reported transverse FWHM of 3.79 mm at 1 cm radial offset.⁶
A digital bismuth-germanate Omni Legend reported average spatial resolution of 3.9 mm FWHM at 1 cm.⁷
A total-body long-axial-FOV uEXPLORER reported spatial resolution of ≤3.0 mm FWHM near the center of its field of view.⁸

Because these are F-18 numbers, they understate the blur a clinician sees with Ga-68 PSMA/DOTATATE or Rb-82 cardiac imaging. A scanner specified at 3.8 mm on F-18 will behave closer to the 4.5 mm of our worked example on Ga-68. This is a central caveat when interpreting spec sheets and when setting expectations for a theranostics or cardiac program.⁵⁶

Clinical Impact

Small-lesion detection and the partial volume effect

Spatial resolution and lesion recovery are inseparable through the partial volume effect: activity from an object smaller than roughly twice the reconstructed FWHM spills into neighboring voxels. The object's peak signal is under-recovered, so a small avid node can measure a falsely low SUV — or fall below the detection threshold entirely. For a scanner at 4.5 mm FWHM on Ga-68, lesions smaller than ~9 mm are increasingly under-quantified. This is why resolution matters clinically far beyond "sharpness," and it connects directly to the PET partial volume effect and SUV quantification.

Radionuclide-specific expectations

F-18 (FDG, and F-18-labeled ligands): the sharpest common tracer; positron range is nearly negligible in a clinical bore, and resolution is limited by detector size and non-collinearity.
Ga-68 (PSMA, DOTATATE): clinically excellent but measurably softer than F-18. Small-lesion SUVs are more affected by partial-volume under-recovery; see Ga-68 PSMA PET imaging.
Rb-82 (cardiac perfusion): the largest positron range of the common clinical tracers, which — combined with its short half-life and count-rate demands — is a defining constraint of Rb-82 cardiac PET/MPI. Myocardial wall assessment is generally forgiving of this blur, but small-structure quantification is not.

For a broader tour of these tracers' emission characteristics, see common PET and RPT isotopes.

Time-of-flight does not fix resolution — it improves noise

A frequent point of confusion: time-of-flight (TOF) does not improve spatial resolution. TOF localizes the annihilation along the line of response, which reduces noise and improves effective sensitivity and contrast recovery, but it does not shrink the FWHM set by detector size, non-collinearity, and positron range. Resolution and TOF are complementary; see time-of-flight PET imaging.

Practical Optimization Tips

Set expectations by radionuclide, not by spec sheet. The NEMA F-18 resolution is a best case. Communicate to referrers that Ga-68 and Rb-82 studies are inherently softer for small structures.⁵
Use radionuclide-appropriate resolution recovery. If the reconstruction offers positron-range or point-spread-function modeling, confirm whether the kernel is generic (F-18) or radionuclide-specific; a generic kernel undercorrects Ga-68.⁴
Mind partial-volume effects in SUV reporting. For sub-centimeter lesions, treat SUV as a floor, not a truth, and consider partial-volume-corrected or peak metrics where validated.
Verify resolution at acceptance and after upgrades as part of NEMA-style testing, and confirm that reconstruction changes (including AI-based methods) have not altered the effective PSF.⁵⁶
Match voxel size to resolution. Reconstructing with voxels far finer than the FWHM adds noise without recovering true resolution; excessively coarse voxels waste the resolution you have.
Leverage sensitivity, not just resolution, for small lesions. Higher-sensitivity systems (digital detectors, long axial FOV) improve small-lesion detection by lowering noise even when FWHM is unchanged.⁸

Regulatory Considerations

Spatial resolution is a standardized, reportable performance parameter, and its measurement is anchored in accreditation and equipment-performance frameworks. The relevant references:

NEMA NU 2-2018 (Performance Measurements of Positron Emission Tomographs) defines how transverse and axial spatial resolution are measured with point sources, enabling apples-to-apples comparison across systems. It is the standard cited in acceptance testing and manufacturer specifications.⁵
PET accreditation (for example, ACR and comparable programs) and the annual medical-physics performance evaluation verify that a scanner meets specification and remains stable over time; resolution is baselined at acceptance and monitored through routine QC and phantom imaging.
Radiopharmaceutical use is governed under NRC (10 CFR Parts 20 and 35) or the equivalent Agreement State program — Florida, Maryland, Virginia, California, Nevada, Pennsylvania, New York, and New Jersey are Agreement States, while Washington DC and Delaware are regulated directly by the NRC. The scanner's performance testing is a physics/accreditation matter; the radioactive material is the licensing matter. For context on the NEMA testing suite, see PET/CT NEMA NU-2 performance testing.

A qualified medical physicist ties these threads together: confirming that measured resolution matches specification, that it is interpreted correctly for the radionuclides in clinical use, and that quantitative reporting accounts for its limits.

Frequently Asked Questions (FAQs)

What limits spatial resolution in PET?

Four physical factors dominate: the finite size of the detector element (about half the crystal width at the center of the field of view), annihilation photon non-collinearity (the two 511 keV photons are not exactly 180° apart), positron range (the distance the positron travels before annihilating), and detector penetration/decoding errors. Reconstruction and sampling add further blurring. These combine in quadrature and set a floor below which finer detectors cannot push resolution.

What is positron range and why does it blur PET images?

A positron does not annihilate at the point of decay. It travels a short, energy-dependent distance through tissue, scattering and losing energy, before it meets an electron and annihilates. PET reconstructs the annihilation location, not the decay location, so the separation between them blurs the image. The higher the positron's energy, the farther it travels and the greater the blur.

Why is Ga-68 or Rb-82 imaging blurrier than F-18?

Positron energy differs sharply between radionuclides. F-18 emits low-energy positrons (0.63 MeV maximum) that travel well under a millimeter on average, while Ga-68 (1.90 MeV) and Rb-82 (3.38 MeV) emit high-energy positrons that travel several millimeters. On the same scanner, the higher positron range of Ga-68 and Rb-82 measurably degrades spatial resolution and small-lesion recovery.

What spatial resolution do modern PET/CT scanners achieve?

Modern digital PET/CT systems measured under NEMA NU 2-2018 with an F-18 point source typically report transverse resolution near 3.8–3.9 mm FWHM at a 1 cm radial offset, and total-body long-axial-FOV systems reach about 3.0 mm near the center. Because the standard uses F-18, these numbers understate the blur seen clinically with higher-energy emitters.

Does non-collinearity depend on scanner size?

Yes. The annihilation photons deviate from a perfect 180° by about 0.5° FWHM because the positron-electron pair retains a small residual momentum. This angular uncertainty translates into a positional blur proportional to the detector-ring diameter, so larger-bore scanners suffer more non-collinearity blur than small-bore or preclinical systems.

Can positron range blurring be corrected?

Partly. Resolution-recovery and positron-range corrections during reconstruction can model the blurring kernel, but many use a generic F-18 model that undercorrects high-energy emitters such as Ga-68. Accurate correction requires a radionuclide-specific range kernel, and residual blur remains because the range distribution is broad and tissue-dependent.

How does spatial resolution affect SUV and small lesions?

Limited resolution causes the partial volume effect: activity from a small lesion spills into surrounding voxels, so a lesion smaller than about twice the FWHM measures a falsely low SUV and can be missed entirely. Better resolution improves both detection of small lesions and the quantitative accuracy of SUV measurements.

Key Takeaways

PET resolution is physics-limited, combining detector size, non-collinearity, positron range, and decoding errors in quadrature — the largest term dominates.¹
Positron range is the radionuclide-dependent term: F-18 (~0.6 mm mean range) is sharp; Ga-68 (~2.9 mm) and Rb-82 (~5.9 mm) are progressively softer.²³
Non-collinearity scales with bore size ( $0.0022 D$ ), contributing ~1.8 mm in an 80 cm clinical scanner.¹
NEMA resolution is measured with F-18 (~3.8–3.9 mm at 1 cm on modern digital systems; ~3.0 mm for total-body), so it understates clinical Ga-68/Rb-82 blur.⁵⁶⁸
Resolution drives the partial volume effect, capping accurate SUV recovery for lesions smaller than about twice the FWHM.
TOF improves noise, not resolution — a distinct and complementary benefit.

Conclusion

The sharpness of a PET image is written into the physics of positron decay before any electronics are involved. Detector size and non-collinearity set a clinical floor near 3–4 mm, and positron range lifts that floor further for high-energy emitters — turning a 3.8 mm F-18 scanner into an effectively ~4.5 mm Ga-68 one. Recognizing this budget is what lets a physicist interpret spec sheets honestly, choose and validate the right reconstruction corrections, and counsel clinical teams on where SUV can be trusted and where small-lesion quantification will fall short. Resolution is not a number to admire on a brochure; it is a physics constraint to design programs around.

How DRPS Can Help

Diagnostic Radiation Physics Services supports PET/CT programs with acceptance testing, NEMA-based performance evaluation, annual physics surveys, and practical guidance on resolution, quantification, and radionuclide-specific expectations. Our physicists help sites interpret what their scanner's measured resolution means for FDG, Ga-68, and Rb-82 imaging, and how it should shape SUV reporting and protocol design. We provide PET/CT and nuclear medicine physics, accreditation support, and medical physicist consulting across our service areas, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware. See our locations or contact us.

Related Resources

References

Moses WW. Fundamental limits of spatial resolution in PET. Nucl Instrum Methods Phys Res A. 2011;648(Supplement 1):S236-S240. doi:10.1016/j.nima.2010.11.092. PubMed
Levin CS, Hoffman EJ. Calculation of positron range and its effect on the fundamental limit of positron emission tomography system spatial resolution. Phys Med Biol. 1999;44(3):781-799. doi:10.1088/0031-9155/44/3/019. PubMed
Conti M, Eriksson L. Physics of pure and non-pure positron emitters for PET: a review and a discussion. EJNMMI Phys. 2016;3(1):8. doi:10.1186/s40658-016-0144-5. PubMed
Carter LM, Kesner AL, Pratt EC, et al. The impact of positron range on PET resolution, evaluated with phantoms and PHITS Monte Carlo simulations for conventional and non-conventional radionuclides. Mol Imaging Biol. 2020;22(1):73-84. doi:10.1007/s11307-019-01337-2. PubMed
National Electrical Manufacturers Association. NEMA NU 2-2018: Performance Measurements of Positron Emission Tomographs (PET). Rosslyn, VA: NEMA; 2018. nema.org
Zeimpekis KG, Kotasidis FA, Huellner M, Nemirovsky A, Kaufmann PA, Treyer V. NEMA NU 2-2018 performance evaluation of a new generation 30-cm axial field-of-view Discovery MI PET/CT. Eur J Nucl Med Mol Imaging. 2022;49(9):3023-3032. doi:10.1007/s00259-022-05751-7. PubMed
Smith RL, Bartley L, O'Callaghan C, Haberska L, Marshall C. NEMA NU 2-2018 evaluation and image quality optimization of a new generation digital 32-cm axial field-of-view Omni Legend PET-CT using a genetic evolutionary algorithm. Biomed Phys Eng Express. 2024;10(2). doi:10.1088/2057-1976/ad286c. PubMed
Spencer BA, Berg E, Schmall JP, et al. Performance evaluation of the uEXPLORER total-body PET/CT scanner based on NEMA NU 2-2018 with additional tests to characterize PET scanners with a long axial field of view. J Nucl Med. 2021;62(6):861-870. doi:10.2967/jnumed.120.250597. PubMed